Silver nanoparticles (AgNPs) can be synthesized from a variety of techniques including physical, chemical and biological routes. They have been widely used as nanomaterials for manufacturing cosmetic and healthcare products, antimicrobial textiles, wound dressings, antitumor drug carriers, etc. due to their excellent antimicrobial properties. Accordingly, AgNPs have gained access into our daily life, and the inevitable human exposure to these nanoparticles has raised concerns about their potential hazards to the environment, health, and safety in recent years. From in vitro cell cultivation tests, AgNPs have been reported to be toxic to several human cell lines including human bronchial epithelial cells, human umbilical vein endothelial cells, red blood cells, human peripheral blood mononuclear cells, immortal human keratinocytes, liver cells, etc. AgNPs induce a dose-, size- and time-dependent cytotoxicity, particularly for those with sizes ≤10 nm. Furthermore, AgNPs can cross the brain blood barrier of mice through the circulation system on the basis of in vivo animal tests. AgNPs tend to accumulate in mice organs such as liver, spleen, kidney and brain following intravenous, intraperitoneal, and intratracheal routes of administration. In this respect, AgNPs are considered a double-edged sword that can eliminate microorganisms but induce cytotoxicity in mammalian cells. This article provides a state-of-the-art review on the synthesis of AgNPs, and their applications in antimicrobial textile fabrics, food packaging films, and wound dressings. Particular attention is paid to the bactericidal activity and cytotoxic effect in mammalian cells.

Graphene, graphene oxide, and reduced graphene oxide have been widely considered as promising candidates for industrial and biomedical applications due to their exceptionally high mechanical stiffness and strength, excellent electrical conductivity, high optical transparency, and good biocompatibility. In this article, we reviewed several techniques that are available for the synthesis of graphene-based nanomaterials, and discussed the biocompatibility and toxicity of such nanomaterials upon exposure to mammalian cells under in vitro and in vivo conditions. Various synthesis strategies have been developed for their fabrication, generating graphene nanomaterials with different chemical and physical properties. As such, their interactions with cells and organs are altered accordingly. Conflicting results relating biocompatibility and cytotoxicity induced by graphene nanomaterials have been reported in the literature. In particular, graphene nanomaterials that are used for in vitro cell culture and in vivo animal models may contain toxic chemical residuals, thereby interfering graphene-cell interactions and complicating interpretation of experimental results. Synthesized techniques, such as liquid phase exfoliation and wet chemical oxidation, often required toxic organic solvents, surfactants, strong acids, and oxidants for exfoliating graphite flakes. Those organic molecules and inorganic impurities that are retained in final graphene products can interact with biological cells and tissues, inducing toxicity or causing cell death eventually. The residual contaminants can cause a higher risk of graphene-induced toxicity in biological cells. This adverse effect may be partly responsible for the discrepancies between various studies in the literature.

No abstract

Accelerating convolutional neural networks has recently received ever-increasing research focus. Among various approaches proposed in the literature, filter pruning has been regarded as a promising solution, which is due to its advantage in significant speedup and memory reduction of both network model and intermediate feature maps. To this end, most approaches tend to prune filters in a layer-wise fixed manner, which is incapable to dynamically recover the previously removed filter, as well as jointly optimize the pruned network across layers. In this paper, we propose a novel global & dynamic pruning (GDP) scheme to prune redundant filters for CNN acceleration. In particular, GDP first globally prunes the unsalient filters across all layers by proposing a global discriminative function based on prior knowledge of filters. Second, it dynamically updates the filter saliency all over the pruned sparse network, and then recover the mistakenly pruned filter, followed by a retraining phase to improve the model accuracy. Specially, we effectively solve the corresponding non-convex optimization problem of the proposed GDP via stochastic gradient descent with greedy alternative updating. Extensive experiments show that, comparing to the state-of-the-art filter pruning methods, the proposed approach achieves superior performance to accelerate several cutting-edge CNNs on the ILSVRC 2012 benchmark.

In advanced nanoscience, there is a strong desire to trap and detect nanoscale objects with high-throughput, single-nanoparticle resolution and high selectivity. Although emerging optical methods have enabled the selective trapping and detection of multiple micrometer-sized objects, it remains a great challenge to extend this functionality to the nanoscale. Here, we report an approach to trap and detect nanoparticles and subwavelength cells at low optical power using a parallel photonic nanojet array produced by assembling microlenses on an optical fiber probe. Benefiting from the subwavelength confinement of the photonic nanojets, tens to hundreds of nanotraps were formed in three dimensions. Backscattering signals were detected in real time with single-nanoparticle resolution and enhancement factors of 10(3)-10(4). Selective trapping of nanoparticles and cells from a particle mixture or human blood solution was demonstrated using the nanojet array. The developed nanojet array is potentially a powerful tool for nanoparticle assembly, biosensing, single-cell analysis, and optical sorting.

Polyvinylidene fluoride (PVDF) and polyvinylidene fluoride-trifluoroethylene (P(VDF-TrFE) with excellent piezoelectricity and good biocompatibility are attractive materials for making functional scaffolds for bone and neural tissue engineering applications. Electrospun PVDF and P(VDF-TrFE) scaffolds can produce electrical charges during mechanical deformation, which can provide necessary stimulation for repairing bone defects and damaged nerve cells. As such, these fibrous mats promote the adhesion, proliferation and differentiation of bone and neural cells on their surfaces. Furthermore, aligned PVDF and P(VDF-TrFE) fibrous mats can enhance neurite growth along the fiber orientation direction. These beneficial effects derive from the formation of electroactive, polar β-phase having piezoelectric properties. Polar β-phase can be induced in the PVDF fibers as a result of the polymer jet stretching and electrical poling during electrospinning. Moreover, the incorporation of TrFE monomer into PVDF can stabilize the β-phase without mechanical stretching or electrical poling. The main drawbacks of electrospinning process for making piezoelectric PVDF-based scaffolds are their small pore sizes and the use of highly toxic organic solvents. The small pore sizes prevent the infiltration of bone and neuronal cells into the scaffolds, leading to the formation of a single cell layer on the scaffold surfaces. Accordingly, modified electrospinning methods such as melt-electrospinning and near-field electrospinning have been explored by the researchers to tackle this issue. This article reviews recent development strategies, achievements and major challenges of electrospun PVDF and P(VDF-TrFE) scaffolds for tissue engineering applications.

Optical methods to manipulate and detect nanoscale objects are highly desired in both nanomaterials and molecular biology fields. Optical tweezers have been used to manipulate objects that range in size from a few hundred nanometres to several micrometres. The emergence of near-field methods that overcome the diffraction limit has enabled the manipulation of objects below 100 nm. A highly free manipulation with signal-enhanced real-time detection, however, remains a challenge for single sub-100-nm nanoparticles or biomolecules. Here we show an approach that uses a photonic nanojet to perform the manipulation and detection of single sub-100-nm objects. With the photonic nanojet generated by a dielectric microlens bound to an optical fibre probe, three-dimensional manipulations were achieved for a single 85-nm fluorescent polystyrene nanoparticle as well as for a plasmid DNA molecule. Backscattering and fluorescent signals were detected with the enhancement factors up to ∼103 and ∼30, respectively. The demonstrated approach provides a potentially powerful tool for nanostructure assembly, biosensing and single-biomolecule studies.

Upconversion fluorescence has triggered extensive efforts in the past decade because of its superior physicochemical features and great potential in biomedical and biophotonic studies. However, practical applications of upconversion fluorescence are often hindered by its relatively low luminescence efficiency (<1%). Here, we employ a living yeast or human cell as a natural bio-microlens to enhance the upconversion fluorescence. The natural bio-microlens, which was stably trapped on a fiber probe, could concentrate the excitation light into a subwavelength region so that the upconversion fluorescence of core-shell NaYF:Yb/Tm nanoparticles was enhanced by 2 orders of magnitude. As a benefit of the fluorescence enhancement, single-cell imaging and real-time detection of the labeled pathogenic bacteria, such as Escherichia coli and Staphylococcus aureus, were successfully achieved in the dark fields. This biocompatible, sensitive, and miniature approach could provide a promising powerful tool for biological imaging, biophotonic sensing, and single-cell analysis.